INDEX

Explanations

years or time periods expressed as mid-decade

oai_token-act-pair · gpt-3.5-turbo

references to the term "mid" followed by a specific decade

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.6.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

edIn

-0.78

 Bearing

-0.71

Magikarp

-0.70

itsch

-0.65

 Immunity

-0.64

éĹĺ

-0.63

 Taste

-0.62

naires

-0.61

 Masquerade

-0.61

 Redemption

-0.61

POSITIVE LOGITS

west

0.96

western

0.95

-'

0.95

season

0.93

century

0.91

wives

0.90

terms

0.89

cellaneous

0.89

iors

0.88

eteen

0.87

Activations Density 0.015%

years or time periods expressed as mid-decade

references to the term "mid" followed by a specific decade

No Comments

No Known Activations